Re: When Did We Give Away the Internet?Mike O'Donnell – Dec 13, 2003 11:08 AM PST
"And who ultimately is in charge of the Domain Name System? You are. You determine what domains you view. Somewhere on your personal computer is a configuration option that determines where you go to resolve top-level domains, ..."
I think it's time to take this point more seriously as a basis for action. Some modest deployment of new names entirely within the current DNS software and authority structure could devalue the ICANN "throne," and let alternatives such as ORSC act as potential models for the future.
Imagine a root with no meaningful TLD names, only meaningless numerical handles, assigned promiscuously to all requesters. Since no individual handle has inherent value, there is little or no conflict over handle claims. Nobody tries to remember these handles. Rather, we embed selected ones in our own software tables, just the way we now embed IP addresses of our chosen name servers. Instead of thinking of top-level handles as horrible domain names, think of them as permanent portable IP numbers.
The owner of a handle may attach it to any service whatsoever, including a 2d-level DNS zone. A natural consensus favoring one particular 2d-level zone may return us to a natural monopoly similar to the current one. But that monopoly is always subject to challenge through the open registration and advertisement of alternatives, without resort to governing bodies.
In order to try out this proposed regime, there's no need to stage a coup and capture the current root. Implement the handle zone at the third level (e.g., handlezone.sponsor.org) of the current DNS. Any sponsor capable of operating a name server for a busy zone could do this (e.g. ORSC). If the service is sufficiently popular, users and coders will gradually promote it by bookmarking, by recognizing it in browsers' URL-completion functions. Eventually, root zones (perhaps starting with ORSC) can actually map the handles as TLDs, but that isn't essential to success.
The best part of this proposal is that there is no serious downside. If it flops, nothing is lost from the current DNS. The handle zone only challenges the DNS root for user attention, not for any resource allocated by ICANN or Verisign.
"And who ultimately is in charge of the Domain Name System? You are. You determine what domains you view. Somewhere on your personal computer is a configuration option that determines where you go to resolve top-level domains, ..."
I think it's time to take this point more seriously as a basis for action. Some modest deployment of new names entirely within the current DNS software and authority structure could devalue the ICANN "throne," and let alternatives such as ORSC act as potential models for the future.
Imagine a root with no meaningful TLD names, only meaningless numerical handles, assigned promiscuously to all requesters. Since no individual handle has inherent value, there is little or no conflict over handle claims. Nobody tries to remember these handles. Rather, we embed selected ones in our own software tables, just the way we now embed IP addresses of our chosen name servers. Instead of thinking of top-level handles as horrible domain names, think of them as permanent portable IP numbers.
The owner of a handle may attach it to any service whatsoever, including a 2d-level DNS zone. A natural consensus favoring one particular 2d-level zone may return us to a natural monopoly similar to the current one. But that monopoly is always subject to challenge through the open registration and advertisement of alternatives, without resort to governing bodies.
In order to try out this proposed regime, there's no need to stage a coup and capture the current root. Implement the handle zone at the third level (e.g., handlezone.sponsor.org) of the current DNS. Any sponsor capable of operating a name server for a busy zone could do this (e.g. ORSC). If the service is sufficiently popular, users and coders will gradually promote it by bookmarking, by recognizing it in browsers' URL-completion functions. Eventually, root zones (perhaps starting with ORSC) can actually map the handles as TLDs, but that isn't essential to success.
The best part of this proposal is that there is no serious downside. If it flops, nothing is lost from the current DNS. The handle zone only challenges the DNS root for user attention, not for any resource allocated by ICANN or Verisign.
Mike O'Donnell
--------------
More info:
Bob Frankston proposes dotDNS.
A variety of draft articles.